J. N. Am. Benthol. Soc., 2009, 28(2):440–453 Ó 2009 by The North American Benthological Society DOI: 10.1899/07-083.1 Published online: 7 April 2009

Does chemistry differentially affect breakdown in tropical vs temperate streams? Importance of standardized analytical techniques to measure leaf chemistry

1 2 Marcelo Ardo´n AND Catherine M. Pringle Odum School of Ecology, University of Georgia, Athens, Georgia 30602 USA

Susan L. Eggert3 Department of Entomology, University of Georgia, Athens, Georgia 30602 USA

Abstract. Comparisons of the effects of leaf litter chemistry on leaf breakdown rates in tropical vs temperate streams are hindered by incompatibility among studies and across sites of analytical methods used to measure leaf chemistry. We used standardized analytical techniques to measure chemistry and breakdown rate of from common riparian species at 2 sites, 1 tropical and 1 temperate, where a relatively large amount of information is available on litter chemistry and breakdown rates in streams (La Selva Biological Station, Costa Rica, and Coweeta Hydrologic Laboratory, North Carolina, USA). We selected 8 common riparian tree species from La Selva and 7 from Coweeta that spanned the range of chemistries of leaf litter naturally entering streams at each site. We predicted that concentrations of secondary compounds would be higher in the tropical species than in the temperate species and that high concentrations of condensed tannins would decrease breakdown rates in both sites. Contrary to our predictions, mean concentration of condensed tannins was significantly greater (2.63, p , 0.001) for species at Coweeta than for species at La Selva. Concentration of condensed tannins was negatively correlated with breakdown rate among Coweeta species (r ¼0.80), not among La Selva species, and negatively correlated when the 2 sites were combined (r ¼0.53). Concentrations of structural compounds were strongly correlated with breakdown rate at both sites (Coweeta species, lignin r ¼0.94, cellulose r ¼0.77; La Selva species, cellulose r ¼0.78, C r ¼0.73). The chemistries of 8 riparian species from La Selva and 7 riparian species from Coweeta were not as different as expected. Our results underline the importance of standardized analytical techniques when making cross-site comparisons of leaf chemistry.

Key words: organic matter, decomposition, tannins, phenolics, lignin, nitrogen, carbon.

Leaf litter is an important energy source for 2) low N content (Melillo et al. 1983), 3) high lignin:N organisms in forested streams (Wallace et al. 1999). and C:N ratios (Enriquez et al. 1993), and 4) high The chemistry of leaf litter affects its decomposition concentrations of tannins (Ostrofsky 1997, Driebe and Whitham 2000). The effect of leaf litter chemistry on rate, persistence, and availability to consumers (Web- breakdown has been studied extensively in temperate ster and Benfield 1986). Leaf chemical characteristics streams, but leaf litter in tropical streams has not known to decrease breakdown rate include: 1) high received as much attention. The physical and biolog- concentrations of lignin (Gessner and Chauvet 1994), ical differences between tropical and temperate

1 streams (Covich 1988) provide good reason to expect Present address: Duke University, Department of that the factors controlling leaf litter breakdown might Biology, Durham, North Carolina 27708 USA. E-mail: differ between these 2 regions. [email protected] 2 E-mail address: [email protected] Leaves of many tropical species tend to have 3 Present address: Northern Research Station, Forest higher concentrations of secondary compounds and Service, US Department of Agriculture,1831 Hwy 169 E, more recalcitrant compounds than do leaves of Grand Rapids, Minnesota 55744 USA. E-mail: seggert@fs. temperate species (Levin 1976, Coley and Aide 1991). fed.us Previous researchers hypothesized that high concen- 440 2009] TROPICAL–TEMPERATE COMPARISON OF LEAF CHEMISTRY 441 trations of secondary compounds (e.g., tannins) in Methods many tropical species inhibit leaf breakdown rates in Site descriptions tropical streams (Stout 1989, Irons et al. 1994). Stout (1989), using primarily published data, hypothesized La Selva Biological Station (lat 108260N, long that high concentrations of condensed tannins in 848010W; 1536 ha) is in the Caribbean lowlands of tropical species decrease leaf breakdown rate in Costa Rica, at the lowland end of the last protected tropical streams. Most studies that have examined this unbroken biological corridor spanning altitudinal hypothesis have incubated tropical leaves in temperate extremes in Central America. La Selva receives 4000 streams (but see Wantzen and Wagner 2006). In a mm of rain annually, with .400 mm/mo from May to temperate Australian stream, condensed tannin con- December (Sanford et al. 1994). Mean annual stream tent was negatively correlated with breakdown rates water temperature ranges from 24 to 268C. Streams at of 6 species of Eucalyptus, but not with breakdown La Selva display a wide range of variation in solute rates of 12 other tropical and temperate species concentrations and pH because of geothermally (Campbell and Fuchshuber 1995). Leaf litter from modified groundwater inputs (Pringle and Triska tropical species with higher concentrations of tannins 1991). Algal assemblages in streams at La Selva are was processed more slowly by shredders in a German light limited by dense canopy cover (.90%). Thus, stream than was leaf litter from temperate species with food webs are detritus based (Pringle et al. 1993). Most lower concentrations of tannins (Wantzen et al. 2002). in La Selva are nondeciduous, and leaf fall is Results from these previous studies might not be an continuous throughout the year (Hartshorn 1983). 0 accurate reflection of differences in leaf litter chemistry Coweeta Hydrologic Laboratory (lat 35800 N, long 0 and leaf breakdown dynamics between tropical and 83830 W; 2185 ha watershed) is a US Department of temperate streams for 3 reasons: 1) literature reviews Agriculture Forest Service research facility and Long are hindered by the use of different analytical Term Ecological Research site in the Blue Ridge techniques to measure secondary compounds across Province of the Southern Appalachian Mountains studies (Stout 1989), 2) chemical changes in leaves (North Carolina). Coweeta Hydrologic Laboratory during decomposition occur faster in tropical than in has 50þ y of hydrologic and climate monitoring data. temperate streams (Ardo´n and Pringle 2008), and 3) Coweeta receives ;1700 mm of rain annually, and the microbial and invertebrate fauna of temperate stream temperatures range from 4 to 178C (Swank and streams (i.e., where tropical leaves were incubated) Crossley 1988). Dissolved concentrations of N and P might not have been adapted to the chemical are usually low (,10 lg/L), and pH averages ;6.6 to characteristics of tropical leaves. The effects of second- 6.8 (Wallace et al. 1986). Streams in Coweeta drain ary compounds (measured with standard analytical mixed-hardwood forests dominated by oaks and techniques) on breakdown rate of native species hickories and are heavily shaded by riparian rhodo- decomposing in their natural habitats at both a tropical dendron (Wallace et al. 1982). Thus, food webs are and temperate site remain to be explored. detritus based (Wallace et al. 1999). The objectives of our study were to: 1) directly compare initial leaf litter chemistry of common native Selection of tree species for chemical analyses and riparian tree species spanning a wide range of quality breakdown experiments from a tropical site (La Selva Biological Station, Costa At least 2015 native plant species exist at La Selva, Rica) and a temperate site (Coweeta Hydrologic but the exact number that can be considered riparian is Laboratory, North Carolina, USA) using standardized unclear (C. Taylor, Missouri Botanical Garden, person- analytical techniques and 2) examine the importance of al communication). Up to 116 plant species have been leaf secondary compounds on leaf breakdown at these reported near streams and wetlands in vegetation 2 sites, where a relatively large amount of information studies at La Selva (Lieberman et al. 1985). A list of exists on plant chemistry and breakdown dynamics in trees and shrubs species that commonly occur along streams (Hartshorn 1983, Swank and Crossley 1988, streams in La Selva was compiled based on the Rosemond et al. 1998, Webster et al. 1999). We literature (Hartshorn 1983, Lieberman et al. 1985, predicted that concentrations of secondary compounds Hartshorn and Hammel 1994) and consultation with (total phenolics, condensed and hydrolyzable tannins) experts on the La Selva flora (B. Hammel and M. would be higher among species from the tropical site Grayum, Missouri Botanical Garden; O. Vargas and N. than among species from the temperate site, and that Zamora, Organization for Tropical Studies and Costa secondary compounds would inhibit breakdown rates Rica’s Instituto Nacional de Biodiversidad; Appendix). in both sites. Species selected from this list for inclusion in the litter 442 M. ARDO´ NETAL. [Volume 28 breakdown experiment were chosen to span the range tannins and hydrolyzable tannins can be higher than of litter chemical quality of leaves that naturally enter total phenolics in our assays because they are La Selva streams. These species had high (Zygia measured with different reactions (Schultz and Bald- longifolia [Fabaceae, sotacaballo], Carapa nicaraguensis win 1982). All samples were compared to standards [Meliaceae, caobilla]), intermediate (Castilla elastica prepared from pooled litter samples from each site to [Moraceae, hule], Luehea seemannii [Malvaceae, gua´ci- avoid problems associated with use of commercial mo colorado], Simira maxonii [Rubiaceae, guatil colo- standards (Appel et al. 2001). Leaf C and N content rado], Terminalia oblonga [Combretaceae, sura]), and were determined using a Carlo Erba NA 1500 CHN low (Trema integerrima [Ulmaceae, capulı´n], Ficus analyzer (Carlo Erba, Milan, Italy). For P analysis, insipida [Moraceae, chilamate]) concentrations of sec- ground leaf material was weighed into acid-washed ondary compounds (Janzen and Waterman 1984, Stout and precombusted ceramic crucibles, combusted at 1989). 5008C, acid-digested, and analyzed spectrophotomet- Approximately 610 plant species exist at Coweeta. rically (ascorbic acid method; APHA 1998). An estimated 27 tree species and 4 shrub species commonly occur alongside streams (mesic, hemlock– Measurements of leaf breakdown rate poplar community; Elliott et al. 1999). Species selected Breakdown rates at La Selva were measured in the for the breakdown experiment were chosen to repre- Sabalo, a 3rd-order stream on the east border of the sent the main categories of leaf breakdown rates in reserve. Water temperature in the Sabalo is fairly Coweeta streams (Wallace et al. 1982, Webster et al. constant (diel range ¼ 24–268C, annual range ¼ 24– 1999). These species had slow (Pinus strobus [Pinaceae, 278C; Rosemond et al. 1998). Daily maximum and white pine], Quercus rubra [Fagaceae, northern red minimum temperatures were recorded with a maxi- oak], Rhododendron maximum [Ericaceae, rhododen- mum–minimum thermometer during the breakdown dron]), medium (Acer rubrum [Aceraceae, red maple], experiment. Plastic mesh bags (20 cm 3 36 cm) were Liquidambar styraciflua [Hamamelidaceae, sweetgum]), filled with 5 g of dried leaf litter. The bags had a coarse or fast (Cornus florida [Cornaceae, dogwood], Lirioden- mesh (5 mm) so that stream invertebrates could access dron tulipifera [Magnoliaceae, tulip poplar]) breakdown the litter (Benfield 1996). Five grams of litter was used rates. to approximate the average mass of naturally occur- Freshly fallen leaves from La Selva were collected in ring leaf packs in streams at La Selva (5.3 6 3g;M. January and February 2003 and 2004. Recently Ardo´n, unpublished data), and this mass of litter was senesced leaves were collected from the ground (leaves used in previous studies at La Selva (Rosemond et al. that had begun to decompose or showed signs of 1998, 2001, 2002, Ardo´n et al. 2006, Stallcup et al. 2006). herbivory were avoided), brought to the laboratory, Litter bags were anchored to the stream bed with and stored in an air conditioned room until used. metal stakes. Litter bags were placed in the stream in Freshly fallen leaves from Coweeta were collected in February 2003 and 2004 and were collected over 7 October 2003 and 2004 as described for leaves from La dates for each species (the dates for each species Selva. depended on their breakdown rate) over a period of 4 mo. Data reported for F. insipida and Z. longifolia, Leaf litter chemical analyses which were used both years, are from the 2003 Standardized analytical methods were used to experiment. On each collection date, 4 litter bags were measure the chemistry of leaves from La Selva and brought to the laboratory. Litter was washed to Coweeta. Structural compounds (i.e., cellulose, hemi- remove insects and sediments, oven dried (408C, 24– cellulose, and lignin) were estimated with sequential 48 h), weighed, combusted (5008C, 1 h), and reweighed neutral detergent/acid detergent digestion on an to obtain ash-free dry mass (AFDM) remaining. Ankom 2000 fiber analyzer (ANKOM Fiber Technol- Chemical changes in the leaves during breakdown ogies, Fairport, New York; Goering and Van Soest and microbial and insect colonization of these leaves 1970). Three separate analyses were conducted for are reported elsewhere (Ardo´n and Pringle 2008). secondary compounds: condensed tannins (acid-buta- Breakdown rates at Coweeta were measured in a 1st- nol assay; Rossiter et al. 1988), hydrolyzable tannins order stream (catchment 53). Daily mean, maximum, (potassium-iodate technique; Schultz and Baldwin and minimum temperatures were recorded using Ryan 1982), and total phenolics (Folin–Denis assay; Swain temp-mentors (Ryan Instruments Inc., Washington). 1979). Samples were extracted in 70% acetone with 1 Plastic mesh bags (20 cm 3 36 cm, 5-mm mesh; Eggert mM ascorbic acid and evaporated under reduced and Wallace 2003) were filled with 15 g of dried leaves. pressure to provide aqueous extracts. Condensed The average mass of naturally occurring leaf packs at 2009] TROPICAL–TEMPERATE COMPARISON OF LEAF CHEMISTRY 443

Coweeta is 15 g (J. B. Wallace, University of Georgia, personal communication) and has been widely used in previous studies at this site (Webster et al. 1999). Litterbags were placed in the stream on 1 December 2003 and 2004, and 3 to 7 replicates were collected on 9 to 11 collection dates (Eggert and Wallace 2003). Litter bags were placed on ice and transported back to the laboratory for processing. Litter was washed to remove invertebrates and sediments, oven dried 0.80 (0.010) 44.9 (0.44) (608C, 7 d), weighed, combusted (5008C, 12 h), and reweighed to obtain AFDM remaining. )

Data analyses Breakdown rate, k, was estimated by linear regres- sion of ln(x)-transformed % AFDM remaining vs day (d) and vs degree day (dd) for both sites (negative exponential model; Benfield 1996). Relationships be- tween breakdown rate (per d and per dd) and concentrations of leaf litter chemical constituents were examined in species from both sites with Pearson correlation coefficients (Proc Corr; SAS; Version 9.1, SAS Institute, Cary, North Carolina). Principal com- ponent analysis (PCA; SAS) on untransformed leaf chemistry data was used to examine overall variation in leaf chemistry between the 2 sites (SAS). Mean concentrations of chemical constituents were com- pared with analysis of variance (ANOVA on ln(x þ 1)- transformed data to meet assumptions of normality; SAS).

Results Leaf litter chemistry tannins Lignin Cellulose Hemicellulose P N C

Leaf chemistry varied widely among species from Hydrolyzable the 2 sites (Table 1). Mean concentration of condensed tannins was higher for Coweeta than for La Selva species (F1,58 ¼ 24.3, p , 0.001; Fig. 1A). Both total phenolics and hydrolyzable tannins were slightly Total higher for Coweeta than for La Selva species (Fig. phenolics 1B, C). Lignin was higher for Coweeta than for La Selva species (F1,58 ¼ 4.27, p , 0.05; Fig. 1D). Cellulose concentrations were similar between both sites, where- tannins 9.3 (0.48) 12.7 (0.97) 14.7 (0.33) 8.4 (0.15) 12.4 (0.07) 9.0 (0.09) 0.161 (0.002 as hemicellulose was higher for La Selva than for 5.5 (0.27) 4.4 (0.41) 3.9 (0.09) 15.90.5 (0.96) (0.11) 18.6 11.8 (0.31)2.8 (2.62) (0.27)0.8 15.8 (0.06) 4.3 (0.54) (0.34)8.1 2.7 (0.25) (0.15) 0.9 0.095 (0.14) (0.0008) 10.0 (0.65) 6.6 (0.55) 8.3 (0.78) 1.24 (0.010) 3.0 19.1 (0.12) (0.67) 15.3 7.8 (1.56) 45.5 (0.42) (0.09) 23.0 (0.30) 14.7 (0.65) 5.5 28.3 (0.68) (0.55) 12.2 14.6 24.6 (0.35) (0.20) (0.80) 0.08 (0.0002) 10.1 0.11 17.5 (0.01) (0.61) (0.41) 1.27 (0.02) 0.10 0.06 35.0 (0.01) (0.001) (0.32) 2.05 (0.07) 1.87 44.9 (0.04) (0.41) 1.61 (0.06) 46.4 (0.27) 34.1 (0.27) 11.3 (2.34) 12.3 (3.72) 9.4 (1.51) 13.4 (0.72) 19.5 (0.83) 22.0 (1.37) 0.16 (0.01) 2.03 (0.10) 40.4 (0.52) 26.8 (1.10) 35.3 (15.20) 39.7 (1.68) 16.5 (0.46) 14.8 (0.24) 9.5 (0.21) 0.032 (0.00001) 0.41 (0.002) 47.6 (0.06) 22.3 (1.63) 6.8 (0.19) 9.6 (0.43) 14.9 (0.29) 13.0 (0.09) 9.3 (0.24) 0.069 (0.0008) 0.58 (0.002) 44.4 (0.06) 57.7 (4.88)35.9 (3.50)24.8 (2.31) 5.6 (0.41) 7.7 (1.64) 25.3 (12.60)23.2 10.2 (2.82) (1.37) 13.1 17.2 (1.68) (1.88) 32.3 (5.94) 19.2 (1.79)13.6 (1.75) 16.9 21.1 (0.21) (0.40) 22.3 (0.66) 34.912.8 (5.32) 21.3 23.6 (3.76) 7.4 (0.34) (0.25) (1.41) 15.6 11.5 (0.15) 18.3 (1.35) (2.36) 16.0 (0.67) 14.0 9.8 (2.30) (0.70) 18.0 (1.41) 0.041 (0.0002) 11.4 (1.00) 0.036 (0.0001) 16.3 0.032 (0.41) (0.001) 0.40 (0.002) 7.5 (1.09) 19.9 0.68 (0.35) (0.003) 6.9 (0.81) 50.3 (0.09) 0.32 (0.004) 47.3 0.08 (0.01) 17.5 (0.0002) (0.20) 15.5 47.3 (0.65) (0.10) 13.9 0.91 (0.64) (0.13) 0.11 (0.01) 41.7 (0.23) 0.12 (0.01) 1.27 (0.07) 1.36 44.2 (0.07) (0.05) 39.2 (0.23) Condensed Coweeta species (F1,58 ¼ 4.19, p , 0.05; Fig. 1E, F). P and N concentrations were higher in La Selva than in Coweeta species (F1,58 ¼ 13.5, p , 0.001; F1,58 ¼ 102.5, p , 0.0001; respectively; Fig. 1G, H). C concentration was higher for Coweeta than for La Selva species (F1,58 ¼ 45.1, p , 0.0001; Fig. 1I).

PCA of leaf chemistry data did not show a clear Species 1. Mean (SE) concentrations (% dry mass) of chemical constituents of leaves from 7 common riparian species in Coweeta Hydrologic Laboratory, North separation of species from the 2 sites (Fig. 2). PC axis 1

explained 40% of the variation and was positively ABLE T Acer rubrum Cornus florida Liquidambar styraciflua Liriodendron tulipifera Pinus strobus Ficus insipida Luehea seemannii Terminalia oblonga Zygia longifolia Range 0.4–23.1 0.9–32.3 2.8–34.8 5.5–28.3 12.2–24. 6 7.5–22.0 0.058–0.160 0.91–2.05 34.1–46.4 Quercus rubra Rhododendron maximum RangeCarapa nicaraguensis Castilla elastica Simira maxonii Trema integerrima 5.5–57.7 4.4–35.3 3.9–39.7 8.4–21.1 12.4–3.6 9.0–16 0.032–0.161 0.32–1.24 44.4–50.3 Coweeta Carolina, and 8 common riparian species from La Selva Biological Station, Costa Rica. correlated with condensed tannins (r ¼ 0.41) and La Selva 444 M. ARDO´ NETAL. [Volume 28

FIG. 1. Mean (61 SE) concentrations of (A) condensed tannins, (B) total phenolics, (C) hydrolyzable tannins, (D) lignin, (E) cellulose, (F) hemicellulose, (G) P, (H) N, and (I) C of leaves from 7 riparian species from Coweeta Hydrologic Laboratory, North Carolina, and 8 riparian species from La Selva Biological Station, Costa Rica. Significant differences are from analysis of variance. * ¼ p , 0.05, ** ¼ p , 0.01, *** ¼ p , 0.001. 2009] TROPICAL–TEMPERATE COMPARISON OF LEAF CHEMISTRY 445

FIG. 2. Principal component analysis (PCA) of leaf chemistry of 7 riparian species from Coweeta Hydrologic Laboratory, North Carolina, and 8 riparian species from La Selva Biological Station, Costa Rica. PC axis 1 explained 40% of the variation, and PC axis 2 explained 27% of the variation. Coweeta species are denoted by black symbols, La Selva species are denoted by white symbols. CT ¼ condensed tannins, hemicell ¼ hemicellulose. Error bars indicate 1 SD.

negatively correlated with P (r ¼0.40) and N (r ¼ 3). Breakdown rates (per dd) of Coweeta species were 0.39) content. PC axis 2 explained 27% of the variation strongly correlated with lignin (r ¼0.94, p , 0.01), and was positively correlated with cellulose (r ¼ 0.52), cellulose (r ¼0.77, p , 0.05), P (r ¼ 0.90, p , 0.01), and hemicellulose (r ¼ 0.50), and lignin (r ¼ 0.37) content. condensed tannins (r ¼0.80, p , 0.05; Table 3). Breakdown rates (per dd) of La Selva species were Leaf breakdown rate strongly correlated with structural compounds (cellu- lose, r ¼0.78, p , 0.05) and %C (r ¼0.73, p , 0.01; Mean breakdown rates were 33 to 73 faster at La Table 3). Concentration of condensed tannins was not Selva (mean k ¼ 0.058/d, 0.0023/dd) than at Coweeta significantly correlated with breakdown rates (per dd) (mean k ¼ 0.008/d, 0.0008/dd; Table 2). The difference in La Selva species (Table 3). When all 15 species were in breakdown rate between the slowest and the fastest analyzed together, breakdown rates (per dd) were species was much higher at La Selva (153 based on k strongly negatively correlated with %C (r ¼0.76, p , [per d], 183 based on k [per dd]) than at Coweeta (53 0.01) and lignin (r ¼0.66, p , 0.01; Table 3) and based on k [per d] and 83 based on k [per dd]; Table 2). marginally significantly negatively correlated with Breakdown rates (per dd) of the 4 slowest decompos- condensed tannins (r ¼0.53, p ¼ 0.04). ing species in each site (i.e., C. elastica, L. seemannii, Z. longifolia, and C. nicaraguensis in La Selva; A. rubrum, Discussion Q. rubra, P. strobus, and R. maximum in Coweeta) were not significantly different (F1,6 ¼ 4.08, p ¼ 0.08; Fig. 3). Contrary to our initial predictions, secondary Overall, relationships between leaf chemical param- compounds were higher in species at the temperate eters and breakdown rates were stronger among (Coweeta) than the tropical (La Selva) site and Coweeta species than among La Selva species (Table concentrations of condensed tannins were correlated 446 M. ARDO´ NETAL. [Volume 28

TABLE 2. Breakdown rates calculated for 7 common nins and phenolics in litter of 8 tropical and 10 riparian species from Coweeta Hydrologic Laboratory, North temperate species in Australia. Their results, combined Carolina, and 8 species from La Selva Biological Station, with ours, bring into question the generalization that Costa Rica. Breakdown rates (k) are classified as slow (S), concentrations of secondary compounds are greater in medium (M), or fast (F) based on days (Webster and Benfield leaf litter from tropical than from temperate species. Of 1986) and on degree days (dd) (Cummins et al. 1989). the species analyzed, temperate P. strobus had the k highest concentration of condensed tannins. Conifers tend to have high concentrations of tannins and Location Species (/d) (/dd) phenolics, which can inhibit breakdown (Ostrofsky Coweeta Cornus floridaa 0.017 F 0.0017 F 1997), and P. strobus is a poor resource for aquatic Liriodendron tulipifera 0.010 F 0.0011 M invertebrates (Whiles and Wallace 1997). Our decision Liquidambar styraciflua 0.008 M 0.0009 S to include P. strobus might have biased our results, but Acer rubrum 0.008 M 0.0008 S Quercus rubra 0.007 M 0.0007 S it is a species representative of poor quality litter that Rhododendron maximum 0.004 S 0.0004 S commonly enters Coweeta streams. Pinus strobus 0.003 S 0.0002 S Concentrations of condensed tannins were more La Selva Trema integerrima 0.198 F 0.0076 F strongly related to breakdown rates of temperate Ficus insipida 0.079 F 0.0031 F Simira maxonii 0.048 F 0.0018 F Coweeta species than of tropical La Selva species, Terminalia oblonga 0.039 F 0.0015 M which also led to a significant correlation when we Castilla elastica 0.038 F 0.0014 M combined both sites (Fig. 1, Table 3). Evidence for an Luehea seemannii 0.033 F 0.0013 M inhibitory role of condensed tannins and other Carapa nicaraguensis 0.023 F 0.0009 S Zygia longifolia 0.011 F 0.0005 S phenolic compounds on leaf breakdown is stronger in temperate than in tropical streams. Ostrofsky (1997) a Previously reported in Wallace et al. (1982) found a significant negative relationship between condensed tannins and leaf breakdown rate for 48 with breakdown rate only for species at Coweeta. temperate species. Gessner and Chauvet (1994) report- When we combined both sites, condensed tannins ed a negative relationship between condensed tannins were correlated with breakdown rate (per dd), but it and leaf breakdown rate of 5 species in a French was mostly driven by the Coweeta species. Use of stream. Hybridization in cottonwood increased standardized analytical techniques and incubation of condensed tannin concentrations and led to slower leaf leaves from native species in their natural habitat breakdown in temperate streams (Driebe and Whith- enabled us to compare leaf chemistry directly and to am 2000). Phenolic compounds in Eucalyptus leaves relate leaf chemistry to leaf breakdown across a affect leaf breakdown by inhibiting fungal colonization tropical and temperate site. In contrast, previous (Ba¨rlocher et al. 1995, Canhoto and Grac¸a 1999) and studies of differences in leaf breakdown between invertebrate feeding (Canhoto and Grac¸a 1999). In temperate and tropical streams have relied on litera- contrast, the few studies conducted with tropical ture values for leaf chemistry (measured in a non- species do not provide evidence for a strong role of standardized manner; Stout 1989), have incubated leaf secondary compounds in leaf breakdown (Campbell litter from tropical species in temperate streams and Fuchshuber 1995, Ardo´n et al. 2006, Ardo´n and (Campbell and Fuchshuber 1995, Wantzen et al. Pringle 2008) and do not support the hypothesis of 2002), or have conducted cross-site transplants in Stout (1989) that condensed tannins inhibit leaf which leaves from a tropical ecosystem were incubat- breakdown of tropical species. ed in a temperate stream and vice versa (Stout 1989, Our results underline the importance of structural Irons et al. 1994, Wantzen and Wagner 2006). compounds in inhibiting breakdown of both tropical and temperate leaves. Concentrations of cellulose and How do secondary compounds differ between a tropical lignin in leaves were strongly correlated with break- and a temperate site? down rate at both sites (Table 3). Specialized enzymes Contrary to our predictions based on the literature are required to process recalcitrant forms of C, such as (Levin 1976, Coley and Aide 1991), concentrations of cellulose and lignin (Melillo et al. 1983). Our results condensed tannins were greater (2.63, p , 0.001) in agree with those of previous studies that have shown temperate Coweeta species than in tropical La Selva that high concentrations of recalcitrant forms of C, species (Fig. 1). In a study similar to ours, Campbell such as lignin, inhibit leaf breakdown (Melillo et al. and Fuchshuber (1995) found no statistically signifi- 1983, Gessner and Chauvet 1994, Hutchens and cant differences in concentrations of condensed tan- Benfield 2000). 2009] TROPICAL–TEMPERATE COMPARISON OF LEAF CHEMISTRY 447

FIG. 3. Breakdown rates (k/degree day [dd]) of leaves from common riparian species from (A) La Selva Biological Station, and (B) Coweeta Hydrologic Laboratory.

How do leaf breakdown rates differ between a temperate range of the processing continuum (i.e., range of and a tropical stream? breakdown rates of leaves entering a stream; Petersen and Cummins 1974) might not be very different Differences in breakdown rates between the slowest between a tropical and a temperate site when and the fastest decomposing species at La Selva were exceptionally fast decomposing species are excluded. much greater (15–183) than the differences between However, it remains to be determined how many slow and fast decomposing species at Coweeta (5–83). species in tropical streams are processed as fast as T. The large difference among tropical species was driven integerrima. Trema integerrima is common along streams mostly by the extremely fast breakdown rate of T. in La Selva and in much of Costa Rica (Hartshorn integerrima. The difference in breakdown rate between 1983) and is a high quality, but transient, resource for the 2nd fastest (F. insipida) and the slowest (Z. longifolia) invertebrates. decomposing species at La Selva was similar (6–73)to On average, breakdown rates tended to be faster in the difference between slow and fast decomposing La Selva. However, breakdown rates did overlap species at Coweeta (5–83). This result suggests that the between the 2 sites. Breakdown rates (per dd) of the 448 M. ARDO´ NETAL. [Volume 28

TABLE 3. Pearson correlation coefficients between break- rate between T. integerrima and Z. longifolia was 153 in down rates (per degree day) and concentrations of chemical the Sabalo study stream (Table 4) and 113 in a 1st-order constituents of leaf litter from 7 riparian species from stream (Ardo´n et al. 2006). Our results agree with Coweeta Hydrologic Laboratory, North Carolina, and 8 previous reports of relatively constant differences in riparian species from La Selva Biological Station, Costa Rica. breakdown rates among species under different stream Bold font indicates statistically significant correlations. * ¼ p , 0.05, ** ¼ p , 0.01. conditions (Webster and Benfield 1986). If all break- down rates measured in La Selva were scaled to be Chemical constituent Coweeta La Selva Both sites similar in magnitude to rates at Coweeta (i.e., decreased by a factor of 3), k [per d] would be 23 C 0.79* 0.73* 0.76* N 0.58 0.03 0.37 faster at La Selva than at Coweeta, but k [per dd] P 0.90** 0.01 0.29 would be very similar between sites (mean k/dd ¼ Lignin 0.94** 0.65 0.66* 0.0008 in La Selva, 0.0008 in Coweeta). This result Cellulose 0.77* 0.78* 0.49 supports the hypothesis that the ranges in breakdown Hemicellulose 0.34 0.37 0.16 Condensed tannins 0.80* 0.60 0.53* rates are similar between tropical and temperate Total phenolics 0.10 0.50 0.37 streams when breakdown rates are expressed in Hydrolyzable tannins 0.06 0.46 0.44 degree days (Irons et al. 1994). C:N 0.75* 0.38 0.47 Differences in breakdown rates between the 2 sites N:P 0.38 0.10 0.07 also could have been confounded by the difference in sizes of the litter bags used at the 2 sites (Coweeta, 15 g; La Selva, 5 g). Breakdown rates in litter bags with 4 slower decomposing species in each site (i.e., C. high initial litter mass are slower than are rates in litter elastica, L. seemannii, Z. longifolia, and C. nicaraguensis in La Selva; A. rubrum, Q. rubra, P. strobus, and R. bags with lower initial mass (Reice 1974, Benfield et al. 1979, Campbell et al. 1994, Ruetz et al. 2006). However, maximum in Coweeta) did not differ significantly (F1,6 ¼ 4.08, p . 0.05). Our results suggest that slow the results from these studies are difficult to generalize. decomposing species in tropical streams can break Reice (1974) found decreasing breakdown rates of litter down at rates similar to those of slow decomposing packs with increasing pack size only in summer and species in temperate streams. Leaves from Z. longifolia not in other seasons. Benfield et al. (1979) reported lost 95% of their mass in 686 d in headwater streams at decreasing breakdown rates with increasing leaf mass, La Selva (Ardo´n et al. 2006), a period that is similar to but their results were confounded by changes in both the 656 d to 95% mass loss reported for R. maximum in leaf pack size and exposure method (packs vs bags of Coweeta headwater streams (Greenwood et al. 2007). different meshes). Campbell et al. (1994) found that Differences in leaf breakdown rates between Cow- breakdown rates were faster in bags with 0.5 g initial eeta and La Selva could have been confounded by the mass than in bags with 3 g initial mass, but breakdown fact that rates were measured in a 3rd-order stream at rates were slowest in bags with 1 g initial mass. Ruetz La Selva and a 1st-order stream at Coweeta. Physical et al. (2006) reported that breakdown rates tended to abrasion caused by greater discharge fluctuations in be slower when leaf mass was higher, but the trend the Sabalo might have led to faster breakdown rates was not statistically significant. These studies illustrate thanwouldhavebeenobservedhadratesbeen that leaf mass in packs or bags affects breakdown rate, measured in a smaller stream. Breakdown rates of 4 but the direction and magnitude of the effect are species measured in the Sabalo stream were 23 to 33 difficult to predict. Ruetz et al. (2006) suggested that faster than rates of the same species measured in a 1st- the size of leaf packs and bags should be carefully order stream at La Selva (Table 4), but the relative selected and that the best approach might be to imitate differences among species were similar in both the natural size of leaf accumulations occurring at each streams. For example, the difference in breakdown site (as was done in our study).

rd st TABLE 4. Breakdown rates (/d) of 4 common riparian species in the Sabalo (3 -order stream) and Saltito 100 (1 -order stream) in La Selva Biological Station. F values are from analysis of covariance (ANCOVA). * ¼ p , 0.05, ** ¼ p , 0.01, *** ¼ p , 0.001.

Species Sabalo Saltito-100 F Source Trema integerrima 0.198 0.067 37.74*** Ardo´n et al. (2006) Ficus insipida 0.079 0.053 22.53*** Stallcup et al. (2006) Castilla elastica 0.038 0.010 117.40*** Ardo´n et al. (2006) Zygia longifolia 0.013 0.006 8.79** Ardo´n et al. (2006) 2009] TROPICAL–TEMPERATE COMPARISON OF LEAF CHEMISTRY 449

Are shredders or microbes differentially affecting leaf al. 2002, Wright and Covich 2005, Crowl et al. 2006). breakdown in these 2 sites? The mesh size of our litter bags probably prevented large macroinvertebrates from entering bags and, thus, Irons et al. (1994) suggested that microbes might be reduced their effects on leaf breakdown in La Selva. more important than shredders in tropical streams, However, exclusion of large macroinvertebrates with whereas shredders might be more important than electricity and meshes of different sizes did not microbes in higher latitude streams. In agreement with significantly alter leaf breakdown rate of F. insipida in this hypothesis, recent studies have reported a paucity the Sabalo (Rosemond et al. 1998). Thus, large of shredders in tropical streams relative to temperate macroinvertebrates probably do not play a major role streams (Dobson et al. 2002, Wantzen et al. 2002, in leaf breakdown at La Selva. Wantzen and Wagner 2006). However, macroinverte- brate shredders do occur in high abundance in some Use of standardized analytical techniques to measure leaf tropical streams (Benstead 1996, Cheshire et al. 2005, chemistry Yule et al. 2009). One proposed hypothesis to explain the difference in shredder abundance between tem- Use of standardized analytical methods to measure perate and tropical streams is high concentrations of leaf chemistry would facilitate future cross-site com- secondary compounds in tropical plant species (Stout parisons. We selected methods that are widely used by 1989, Dobson et al. 2002, Wantzen et al. 2002). Our terrestrial ecologists and that are commonly regarded results do not support this hypothesis. Concentrations to be the most accurate. The forage fiber technique was of secondary compounds were higher in leaves of used to measure structural compounds because this species at Coweeta than of species at La Selva, but technique is simpler and more accurate than other shredders were more prevalent at Coweeta than at La commonly used techniques (Ryan et al. 1990). Stan- Selva. Higher abundance (range ¼ 75–164 individuals/ dard colorimetric assays with standards made from litter bag) and biomass (range ¼ 6.4–84 mg/litter bag) purified tannins from our own samples (i.e., self- of macroinvertebrate shredders occur in Coweeta standards; Appel et al. 2001) were used to measure streams (Wallace et al. 1982, Whiles and Wallace secondary compounds. Most studies conducted in 1997, Eggert and Wallace 2003, Greenwood et al. streams have used commercial tannic acid as a 2007) than in La Selva streams (abundance 0–2 standard, despite recognition in the terrestrial litera- individuals/litter bag, biomass 0–0.65 mg/litterbag; ture that commercial tannic acid can overestimate or Rosemond et al. 1998, Ardo´n et al. 2006, Ardo´n and underestimate concentrations by up to a factor of 2 Pringle 2008) (based on the same methods used in our (Hagerman and Butler 1989, Appel et al. 2001). study). Colorimetric methods (after acid-digestion of leaf Higher water temperatures in tropical than in material) can provide accurate and comparable results temperate streams might lead to greater microbial for both N and P (Flindt and Lillebo 2005). activity in tropical than in temperate streams and In conclusion, our results suggest that leaf chemistry might accelerate breakdown rates in tropical streams. is not as different as expected among these 8 riparian Fungal biomass on submerged leaves is higher at La species from La Selva and 7 riparian species from Selva (range ¼ 10–200 mg/g leaf AFDM; Ardo´n and Coweeta (Stout 1989, Wantzen et al. 2002). Contrary to Pringle 2008) than at Coweeta (range ¼ 5–35 mg/g leaf expectations, concentrations of secondary compounds AFDM; Gulis and Suberkropp 2003). Bacterial biomass were correlated with breakdown rates of Coweeta on submerged leaves in La Selva also is high (range ¼ species but not of La Selva species. The wider range of 0.01–1 mg C/g AFDM; Ardo´n and Pringle 2008) and is breakdown rates of La Selva than of Coweeta species similar to that observed on A. rubrum leaves in an was mainly a consequence of the very fast decompo- artificially nutrient-enriched stream in Coweeta (range sition rate of T. integerrima, and comparisons based on ¼ 0.1–0.9 mg C/g AFDM; Gulis and Suberkropp 2003). the other species suggest that the variation in These differences in microbial and invertebrate bio- breakdown rates is not very different between these masses support the hypothesis that microbial process- 2 sites. However, this conclusion should be interpreted ing might be more important in tropical than in with caution because we sampled only a small fraction temperate streams, whereas invertebrate shredders of the possible species entering La Selva and Coweeta might be more important in temperate than in tropical streams. La Selva has .2000 species in an area ½ the streams (Irons et al. 1994). size of Coweeta. Thus, the detritus base of La Selva Recent studies have focused on the effects of larger streams appears to be more species rich than that of macroinvertebrates (shrimp and crabs) on breakdown Coweeta streams. Future studies should examine the rates in tropical streams (March et al. 2001, Dobson et chemical diversity of the detritus base of tropical and 450 M. ARDO´ NETAL. [Volume 28 temperate streams with standardized analytical tech- condensed tannins, and processing rates of tropical and niques to facilitate comparisons of leaf litter chemistry. temperate leaves in an Australian stream. Journal of the North American Benthological Society 14:174–182. Acknowledgements CANHOTO, C., AND M. A. S. GRAC¸A. 1999. Leaf barriers to fungal colonization and shredders (Tipula lateralis) We thank M. D. Hunter for permitting us to run leaf consumption of decomposing Eucalyptus globulus. Mi- chemical analyses in his laboratory. M. Hidalgo crobial Ecology 37:163–172. provided invaluable help with field work in La Selva. CHESHIRE, K., L. BOYERO, AND R. G. PEARSON. 2005. Food webs D. Batzer, A. Covich, E. S. Lindquist, J. L. Meyer, A. D. in tropical Australian streams: shredders are not scarce. Rosemond, J. B. Wallace, L. Boyero, P. Silver, and 2 Freshwater Biology 50:748–769. COLEY,P.D.,AND T. M. AIDE. 1991. Comparisons of herbivory anonymous referees provided helpful suggestions on and plant defenses in temperate and tropical broad- earlier versions of this manuscript. This research was leaved forests. Pages 25–49 in P. W. Pr ice , T. M. supported by National Science Foundation (NSF) Lewinsohn, G. W. Fernandez, and W. W. Benson grants DEB-0075339, DEB-0545463, OISE-0434503, (editors). Plant–animal interactions: evolutionary ecolo- and a University of Georgia Dissertation Completion gy in tropical and temperate regions. John Wiley and Award. We also acknowledge support from the NSF Sons, West Sussex, UK. Coweeta Long-Term Ecological Research project DEB- COVICH, A. P. 1988. Geographical and historical comparisons 0218001 and DEB-0212315. of neotropical streams: biotic diversity and detrital processing in highly variable habitats. Journal of the Literature Cited North American Benthological Society 7:361–386. CROWL, T. A., V. WELSH,T.HEARTSILL-SCALLEY, AND A. P. APHA (AMERICAN PUBLIC HEALTH ASSOCIATION). 1998. Standard COVICH. 2006. Effects of different types of conditioning on methods for the examination of water and wastewater. rates of leaf-litter shredding by Xiphocaris elongate, a 20th edition. American Public Health Association, Amer- Neotropical freshwater shrimp. Journal of the North ican Water Works Association, and Water Environment American Benthological Society 25:198–208. Federation, Washington, DC. CUMMINS, K. W., M. A. WILZBACH,D.M.GATES,J.B.PERRY, AND APPEL, H. M., H. L. GOVERNOR, M. D’ASCENZO,E.SISKA, AND J. W. B. TALIFERRO. 1989. Shredders and riparian vegetation: C. SCHULTZ. 2001. Limitation of foliar phenolics in leaf litter that falls into streams influences communities ecological studies. Journal of Chemical Ecology 27:761– of stream invertebrates. BioScience 39:24–30. 778. DOBSON, M., A. MAGANA,J.M.MATHOOKO, AND F. K. NDEGWA. ARDO´ N,M.,AND C. M. PRINGLE. 2008. Do secondary 2002. Detritivores in Kenyan highland streams: more compounds inhibit microbial- and insect-mediated leaf evidence for the paucity of shredders in the tropics. breakdown in a tropical rainforest stream, Costa Rica? Freshwater Biology 47:909–919. Oecologia (Berlin) 155:311–323. DRIEBE,E.M.,AND T. G. WHITHAM. 2000. Cottonwood ARDO´ N, M., L. A. STALLCUP, AND C. M. PRINGLE. 2006. Does leaf hybridization affects tannin and nitrogen content of leaf quality mediate the stimulation of leaf breakdown by litter and alters decomposition. Oecologia (Berlin) 123: phosphorus in Neotropical streams? Freshwater Biology 99–107. 51:618–633. EGGERT, S. L., AND J. B. WALLACE. 2003. Litter breakdown and BA¨ RLOCHER, F., C. CANHOTO, AND M. A. S. GRAC¸A. 1995. Fungal invertebrate detritivores in a resource-depleted Appala- colonization of alder and eucalypt leaves in two streams chian stream. Archiv fu¨ r Hydrobiologie 156:315–338. in Central Portugal. Archiv fu¨ r Hydrobiologie 133:457– ELLIOTT, K. J., J. M. VOSE, AND W. T. SWANK. 1999. Long-term 470. patterns in vegetation-site relationships in a southern BENFIELD, E. F. 1996. Leaf breakdown in stream ecosystems. Appalachian forest. Journal of the Torrey Botanical Pages 579–589 in F. R. Hauer and G. A. Lamberti Society 126:320–334. (editors). Methods in stream ecology. Academic Press, ENRIQUEZ, S., C. M. DUARTE, AND K. SAND-JANSEN. 1993. San Diego, California. Patterns in decomposition rates among photosynthetic BENFIELD, E. F., R. W. PAUL, AND J. R. WEBSTER. 1979. Influence organisms: the importance of detritus C:N:P content. of exposure technique on leaf breakdown rates in Oecologia (Berlin) 94:457–471. streams. Oikos 33:386–391. FLINDT, M. R., AND A. I. LILLEBO. 2005. Determination of total BENSTEAD, J. 1996. Macroinvertebrates and the processing of nitrogen and phosphorus in leaf litter. Pages 53–59 in M. leaf litter in a tropical stream. Biotropica 28:367–375. A. S. Grac¸a, F. Ba¨rlocher, and M. O. Gessner (editors). CAMPBELL, I. C., G. M. ENIERGA, AND L. FUCHSHUBER. 1994. The Methods to study litter decomposition: a practical guide. influence of pack size, leaf type, and shredder access on Springer Verlag, Dordrecht, the Netherlands. the processing rate of Atherosperma moschatum leaves in GESSNER, M. O., AND E. CHAUVET. 1994. Importance of stream an Australian cool temperate rainforest stream. Inter- microfungi in controlling breakdown rates of leaf litter. nationale Revue der Gesamten Hydrobiologie 79:557– Ecology 75:1807–1817. 568. GOERING, H. K., AND P. J. VAN SOEST. 1970. Forage fiber analysis CAMPBELL,I.C.,AND L. FUCHSHUBER. 1995. Polyphenols, (apparatus, reagents, procedures and some applications). 2009] TROPICAL–TEMPERATE COMPARISON OF LEAF CHEMISTRY 451

Agriculture Handbook No. 379. Agriculture Research in surface waters draining the Atlantic slope of Costa Service, US Department of Agriculture, Washington, DC. Rica. Limnology and Oceanography 38:753–774. GREENWOOD, J. L., A. D. ROSEMOND,J.B.WALLACE,W.F.CROSS, PRINGLE, C. M., AND F. J. TRISKA. 1991. Effects of geothermal AND H. S. WEYERS. 2007. Nutrients stimulate leaf groundwater on nutrient dynamics of a lowland Costa breakdown rates and detritivores biomass: bottom-up Rican stream. Ecology 72:951–965. effects via heterotrophic pathways. Oecologia (Berlin) REICE,S.R.1974.Environmentalpatchinessandthe 151:637–649. breakdown of leaf litter in a woodland stream. Ecology GULIS,V.,AND K. SUBERKROPP. 2003. Leaf litter decomposition 55:1271–1282. and microbial activity in nutrient-enriched and unaltered ROSEMOND, A. D., C. M. PRINGLE, AND A. RAMI´REZ. 1998. reaches of a headwater stream. Freshwater Biology 48: Macroconsumer effects on insect detritivores and detri- 123–134. tus processing in a tropical stream. Freshwater Biology HAGERMAN,A.E.,AND L. G. BUTLER.1989.Choosing 39:515–523. appropriate methods and standards for assaying tannin. ROSEMOND, A. D., C. M. PRINGLE,A.RAMI´REZ, AND M. J. PAUL. Journal of Chemical Ecology 15:1795–1810. 2001. A test of top-down and bottom-up control in a HARTSHORN, G. 1983. Plants. Pages 118–183 in D. H. Janzen detritus-based food web. Ecology 82:2279–2293. (editor). Costa Rican natural history. University of ROSEMOND, A. D., C. M. PRINGLE,A.RAMI´REZ,M.J.PAUL, AND J. Chicago Press, Chicago, Illinois. L. MEYER. 2002. Landscape variation in phosphorus HARTSHORN, G., AND B. E. HAMMEL. 1994. Vegetation types and concentration and effects on detritus-based tropical floristic patterns. Pages 73–89 in L. A. McDade, K. S. streams. Limnology and Oceanography 47:278–289. Bawa, H. A. Hespenheide, and G. S. Hartshorn (editors). ROSSITER, M. C., J. C. SCHULTZ, AND I. T. BALDWIN. 1988. La Selva: ecology and natural history of a Neotropical Relationship among defoliation, red oak phenolics, and rain forest. University of Chicago Press, Chicago, Illinois. gypsy moth growth and reproduction. Ecology 69:267– HUTCHENS, J. J., AND E. F. BENFIELD. 2000. Effects of forest 277. defoliation by the gypsy moth on detritus processing in RUETZ, C. R., M. J. BREEN, AND D. L. VANHAITSMA. 2006. Habitat Southern Appalachian streams. American Midland structure and fish predation: effects on invertebrate Naturalist 143:397–404. colonization and breakdown of stream leaf packs. IRONS, J. G., M. W. OSWOOD,R.J.STOUT, AND C. M. PRINGLE. Freshwater Biology 51:797–806. 1994. Latitudinal patterns in leaf litter breakdown: is RYAN, M. G., J. M. MELILLO, AND A. RICCA. 1990. A comparison temperature really important? Freshwater Biology 32: of methods for determining proximate carbon fractions 401–411. of forest litter. Canadian Journal of Forest Research 20: JANZEN, D. H., AND P. G. WATERMAN. 1984. A seasonal census 166–171. of phenolics, fibre and alkaloids in foliage of forest trees SANFORD, R. L., P. PAABY,J.C.LUVALL, AND E. PHILLIPS. 1994. in Costa Rica: some factors influencing their distribution Climate, geomorphology, and aquatic systems. Pages 19– and relation to host selection by Sphingidae and 33 in L. A. McDade, K. S. Bawa, H. A. Hespenhedie, and Saturniidae. Biological Journal of the Linnean Society G. S. Hartshorn (editors). La Selva: ecology and natural 21:439–454. history of a Neotropical rainforest. University of Chicago LEVIN, D. A. 1976. Alkaloid-bearing plants: an ecogeographic Press, Chicago, Illinois. perspective. American Naturalist 110:261–284. SCHULTZ, J. C., AND I. T. BALDWIN. 1982. Oak leaf quality LIEBERMAN, M., D. LIEBERMAN,G.S.HARTSHORN, AND R. PERALTA. declines in response to defoliation by gypsy moth larvae. 1985. Small-scale altitudinal variation in lowland wet Science 217:149–151. tropical forest vegetation. Journal of Ecology 73:505–516. STALLCUP, L. A., M. ARDO´ N, AND C. M. PRINGLE. 2006. Does MARCH, J. G., J. P. BENSTEAD,C.M.PRINGLE, AND M. W. RUEBEL. nitrogen become limiting under high-P conditions in 2001. Linking shrimp assemblages with rates of detrital detritus-based tropical streams? Freshwater Biology 51: processing along an elevational gradient in a tropical 1515–1526. stream. Canadian Journal of Fisheries and Aquatic STOUT, R. J. 1989. Effects of condensed tannins on leaf Sciences 58:470–478. processing in mid-latitude and tropical streams: a MELILLO, J. M., R. J. NAIMAN,J.D.ABER, AND K. N. ESHLEMAN. theoretical approach. Canadian Journal of Fisheries and 1983. The influence of substrate quality and stream size Aquatic Sciences 46:1097–1106. on wood decomposition dynamics. Oecologia (Berlin) 58: SWAIN, T. 1979. The importance of flavonoids and related 281–285. compounds in fern and ecology. Bulletin of OSTROFSKY, M. L. 1997. Relationship between chemical the Torrey Botanical Club 107:113–153. characteristics of autumn-shed leaves and aquatic SWANK,W.T.,AND D. A. J. CROSSLEY. 1988. Forest hydrology processing rates. Journal of the North American Bentho- and ecology at Coweeta. Springer-Verlag, Berlin, Ger- logical Society 16:750–759. many. PETERSEN, R. C., AND K. W. CUMMINS. 1974. Leaf processing in a WALLACE, J. B., S. L. EGGERT,J.L.MEYER, AND J. R. WEBSTER. woodland stream. Freshwater Biology 4:345–368. 1999. Effects of resource limitation on a detrital-based PRINGLE, C. M., G. L. ROWE,F.J.TRISKA,J.F.FERNANDEZ, AND J. ecosystem. Ecological Monographs 69:409–442. WEST. 1993. Landscape linkages between geothermal WALLACE, J. B., D. S. VOGEL, AND T. F. CUFFNEY. 1986. Recovery activity and solute composition and ecological response of a headwater stream from an insecticide-induced 452 M. ARDO´ NETAL. [Volume 28

community disturbance. Journal of the North American L. TANK,J.J.HUTCHENS, AND D. J. D’ANGELO. 1999. What Benthological Society 5:115–126. happens to allochthonous material that falls into WALLACE, J. B., J. R. WEBSTER, AND T. F. CUFFNEY. 1982. Stream streams? A synthesis of new and published information detritus dynamics: regulation by invertebrate consum- from Coweeta. Freshwater Biology 41:687–705. ers. Oecologia (Berlin) 53:197–200. WHILES,M.R.,AND J. B. WALLACE. 1997. Leaf litter WANTZEN, K. M., AND R. WAGNER. 2006. Detritus processing by decomposition and macroinvertebrate communities in invertebrate shredders: a Neotropical-temperate com- headwater streams draining pine and hardwood catch- parison. Journal of the North American Benthological ments. Hydrobiologia 353:107–119. Society 25:216–232. WRIGHT,M.S.,AND A. P. COVICH. 2005. The effect of WANTZEN, K. M., R. WAGNER,R.SUETFELD, AND W. J. JUNK. 2002. macroinvertebrate exclusion on leaf breakdown rates in How do plant-herbivore interactions of trees influence a tropical headwater stream. Biotropica 37:403–408. coarse detritus processing by shredders in aquatic YULE, C. M., M. Y. LEONG,K.C.LIEW,L.RATNARAJAH,K. ecosystems of different latitudes? Verhandlungen der SCHMIDT,H.M.WONG,R.G.PEARSON, AND L. BOYERO. 2009. Internationalen Vereinigung fu¨ r theoretische und ange- Shredders in Malaysia: abundance and richness are wandte Limnologie 28:815–821. higher in cool upland tropical streams. Journal of the WEBSTER, J. R., AND E. F. BENFIELD. 1986. North American Benthological Society 28:404–415. breakdown in freshwater ecosystems. Annual Review of Ecology and Systematics 17:567–594. Received: 19 July 2007 WEBSTER, J. R., E. F. BENFIELD,T.P.EHRMAN,M.A.SCHAEFFER,J. Accepted: 16 January 2009 2009] TROPICAL–TEMPERATE COMPARISON OF LEAF CHEMISTRY 453

APPENDIX. Common species found along rivers, streams, APPENDIX. Continued. and wetlands in La Selva Biological Station, compiled from the literature (Lieberman et al. 1985, Hartshorn and Hammel Species 1994), and expert opinion of O. Vargas (Organization for Tropical Studies), N. Zamora (INBio), M. Grayum (Missouri Lonchocarpus oliganthus F.J. Herm (Fabaceae) Botanical Garden), and B. Hammel (Missouri Botanical Malvaviscus arboreus Cav. (Malvaceae) Garden). Species denoted with an asterisk were used for Miconia impetiolaris (Sw.) D. Don ex stream leaf breakdown experiments. DC (Melastomataceae) Miconia affinis DC (Melastomataceae) Myriocarpa longipes Species Liebm. (Urticaceae) Neea laetevirens Standl. (Nyctaginaceae) Emergent and canopy tree species restricted to large rivers Ocotea cernua (Nees) Mez (Lauraceae) *Ficus insipida Willd. (Moraceae) Parathesis microcalyx Donn. Sm. (Myrsinaceae) *Zygia longifolia (Humb. and Bonpl. Ex Willd.) Piper auritum Kunth (Piperaceae) Britton and Rose (Fabaceae) Piper biolleyi C. DC. (Piperaceae) Ceiba pentandra L. Gaertn. (Malvaceae) Piper decurrens C. DC. (Piperaceae) Cordia lucidula I.M. Johnston (Boraginaceae) Piper friedrichsthalii C. DC. (Piperaceae) Inga marginata Willd. (Fabaceae) Piper reticulatum L. (Piperaceae) Nectandra reticulata Ruı´z and Pavo´n (Lauraceae) Pleuranthodendron lindenii (Turcz.) Sleumer () Posoqueria latifolia (Rudge) Roem. and Schult (Rubiaceae) Psychotria chagrensis Standley (Rubiaceae) Species of emergent trees common along rivers, Psychotria graciliflora Benth. (Rubiaceae) streams, and swamps Psychotria psychotriifolia (Seem.) Standl. (Rubiaceae) Apeiba membranacea Spruce ex Benth (Tiliaceae) *Simira maxonii (Standl.) Steyerm. (Rubiaceae) *Carapa nicaraguensis C. DC. (Meliaceae) Urera laciniata (Goudot) Wedd. (Urticaceae) *Castilla elastica Sesse´ ex Cerv. (Moraceae) Cecropia obtusifolia Bertol. (Cecropiaceae) Clusia minor L. (Clusiaceae) Goethalsia meiantha (Donn. Sm.) Burret (Tiliaceae) didymantha Donn. Sm. () Iriartea deltoidea Ruı´z and Pavo´n () Inga ruiziana G. Don. (Fabaceae) *Luehea seemannii Triana and Planchon (Tiliaceae) Myrcia splendens (Sw.) DC. (Myrtaceae) novogranatensis Moldenke () Pachira aquatica Aublet (Bombacaceae) (Willd.) Kuntze (Fabaceae) Protium panamense (Rose) I.M. Johnst. (Burseraceae) Pterocarpus officinalis Jacq. (Fabaceae) Socratea exorrhiza (C. Martius) Wendland *Trema integerrima (Beurl) Standl (Ulmaceae) *Terminalia oblonga (Ruiz and Pav.) Steud. (Combretaceae) Terminalia bucidoides Standl. and L.O. Williams (Combretaceae) Virola koschnyi Warb. (Myristicaceae) Welfia georgii Wenland ex Burret (Arecaceae) Subcanopy trees and shrubs common along rivers, streams, and swamps Acalypha diversifolia Jacq. (Euphorbiaceae) Adelia triloba (Muell. Arg.) Hemsl. (Euphorbiaceae) alatum H.F. Loomis (Arecaceae) Bactris longiseta Wendland ex Hemsley (Arecaceae) Carludovica rotundifolia Wendland ex Hooker F. (Cyclanthaceae) Calyptrogyne ghiesbreghtiana (Linden and H. Wendl.) H. Wendl. (Arecaceae) Colubrina spinosa Donn. Sm. (Rhamnaceae) Coussarea impetiolaris Donn. Sm. (Rubiaceae) Cuphea epilobiifolia Koehne (Lythraceae) Cuphea hyssopifolia Kunth (Lythraceae) Cuphea utriculosa Koehne (Lythraceae) Chione costaricensis Standley (Rubiaceae) Cyathea bicrenata Liebm. (Cyatheaceae) L. () Hoffmannia valerioi Standl. (Fabaceae)